We present a comprehensive molecular dynamics (MD) simulation study of composition–structure trends in a set of 25 glasses of widely spanning compositions from the following four systems of increasing complexity: Na₂O–B₂O₃, Na₂O–B₂O₃–SiO₂, Na₂O–CaO–SiO₂–P₂O₅, and Na₂O–CaO–B₂O₃–SiO₂–P₂O₅. The simulations involved new B–O and P–O potential parameters developed within the polarizable shell-model framework, thereby combining the beneficial features of an overall high accuracy and excellent transferability among different glass systems and compositions: this was confirmed by the good accordance with experimental data on the relative BO₃/BO₄ populations in borate and boro(phospho)silicate networks, as well as with the orthophosphate fractions in bioactive (boro)phosphosilicate glasses, which is believed to strongly influence their bone-bonding properties. The bearing of the simulated melt-cooling rate on the borate/phosphate speciations is discussed. Each local {BO₃, BO₄, SiO₄, PO₄} coordination environment remained independent of the precise set of co-existing network formers, while all trends observed in bond-lengths/angles mainly reflected the glass-network polymerization, i.e., the relative amounts of bridging oxygen (BO) and non-bridging oxygen (NBO) species. The structural roles of the Na⁺/Ca²⁺ cations were also probed, targeting their local coordination environments and their relative preferences to associate with the various borate, silicate, and phosphate moieties. We evaluate and discuss the common classification of alkali/alkaline-earth metal ions as charge-compensators of either BO₄ tetrahedra or NBO anions in borosilicate glasses, also encompassing the less explored NBO-rich regime: the Na⁺/Ca²⁺ cations mainly associate with BO/NBO species of SiO₄/BO₃ groups, with significant relative Na–BO₄ contacts only observed in B-rich glass networks devoid of NBO species, whereas NBO-rich glass networks also reveal substantial amounts of NBO-bearing BO₄ tetrahedra.

Body-fluid-exposed bioactive glasses (BGs) integrate with living tissues due to the formation of a biomimetic surface layer of calcium hydroxy-carbonate apatite (HCA) with a close composition to bone mineral. Vast efforts have been spent to understand the mechanisms underlying in vitro apatite mineralization, as either formed by direct precipitation from supersaturated solutions, or from BG substrates in a simulated body fluid (SBF). Formally, these two scenarios are distinct and have hitherto been discussed as such. Herein, we contrast them and identify several shared features. We monitored the formation of amorphous calcium phosphate (ACP) and its crystallization into HCA from a Na ${}_2$ O–CaO–SiO ${}_2$ –P ${}_2$ O ${}_5$ glass exposed to SBF for variable periods out to 28 days. The HCA growth was assessed semi-quantitatively by Fourier transform infrared spectroscopy and powder X-ray diffraction, with the evolution of the relative apatite content for increasing SBF-exposure periods evaluated against trends in Ca and P concentrations in the accompanying solutions. This revealed a sigmoidal apatite growth behavior, well-known to apply to spontaneously precipitated apatite. The results are discussed in relation to the prevailing mechanism proposed for in vitro HCA formation from silicate-based BGs, where we highlight largely simultaneous growth processes of ACP and HCA.

We present a comprehensive study of the Na environments in a large ensemble of 32 silicate-based glass compositions from the borosilicate, phosphosilicate, and borophosphosilicate systems, which comprise either Na as sole glass-network modifier or mixed with Ca. We examined the spatial distribution of Na in the glasses using ²³Na NMR. The relative propensities of Na to associate with the BO₃ and BO₄ structural moieties in B-bearing glasses were probed by heteronuclear dipolar-based ¹¹B{²³Na} magic-angle-spinning NMR experimentation, which yielded both dipolar second moments M₂(B[p]--Na) and M₂(Na--B[p]) for each B[3] and B[4] coordination in a single experiment. These data agreed well with results from atomistic molecular dynamics simulations. Both the spatial distribution of Na and the relative preferences for B[3]--Na and B[4]--Na contacts depend primarily on the amount of non-bridging oxygen (NBO) anions in the glass network, and thereby on the modifier (Na⁺ and Ca²⁺) concentrations, where two regimes were identified: (I) For low modifier contents, the Na⁺ cations are relatively uniformly dispersed across the structure, while there is a strong preference for B^[4]--Na associations. (II) For moderately high modifier contents, Na⁺ distributes randomly and with non-preferential associations with the BO₃ or BO₄ species. However, when the Na⁺ and Ca²⁺ contents are increased further, the growing NBO populations of the glass network coupled with the strong preferential NBO-accommodation at the BO₃ moieties (relative to BO₄) progressively elevates the propensity for B^{[3]}--Na contacts. We discuss the partitioning of the Na reservoir among the BO₃ and BO₄ groups for each regime I and II. We also rationalize the increased disorder of the Na dispersion and the concomitant shift from a preference for B[4]--Na contacts to one for B^{[3]}--Na. Both the nature of the spatial distribution of Na and the relative preferences for B[4]--Na and B[3]--Na contacts in the glass are essentially independent on its precise combination of (B, Si, P) network formers.

We present a solid-state nuclear magnetic resonance (NMR) spectroscopy study of the local ${}^{31}$ P and ${}^1$ H environments in monetite [CaHPO ${}_4$ ; dicalcium phosphate anhydrous (DCPA)], as well as their relative spatial proximities. Each of the three ${}^1$ H NMR peaks was unambiguously assigned to its respective crystallographically unique H site of monetite, while their pairwise spatial proximities were probed by homonuclear ${}^1$ H– ${}^1$ H double quantum–single quantum NMR experimentation under fast magic-angle spinning (MAS) of 66 kHz. We also examined the relative ${}^1$ H– ${}^{31}$ P proximities among the inequivalent {P1, P2} and {H1, H2, H3} sites in monetite; the corresponding shortest internuclear ${}^1$ H– ${}^{31}$ P distances accorded well with those of a previous neutron diffraction study. The NMR results from the monetite phase were also contrasted with those observed from the monetite component present in a pyrophosphate-bearing calcium phosphate cement, demonstrating that while the latter represents a disordered form of monetite, it shares all essential local features of the monetite structure.